[PATCH 1/2] NET: Accurate packet scheduling for ATM/ADSL (kernel)

2006-06-14 Thread Jesper Dangaard Brouer
and how to use it can be found here: http://www.stuart.id.au/russell/files/tc/tc-atm/ A earlier version of the patch, and a _detailed_ empirical investigation of its effects can be found here: http://www.adsl-optimizer.dk/ Signed-off-by: Jesper Dangaard Brouer [EMAIL PROTECTED] Signed-off

[PATCH 2/2] NET: Accurate packet scheduling for ATM/ADSL (userspace)

2006-06-14 Thread Jesper Dangaard Brouer
/files/tc/tc-atm/ A earlier version of the patch, and a _detailed_ empirical investigation of its effects can be found here: http://www.adsl-optimizer.dk/ Signed-off-by: Jesper Dangaard Brouer [EMAIL PROTECTED] Signed-off-by: Russell Stuart [EMAIL PROTECTED] --- diff -Nurp iproute2.orig/include

[PATCH 0/2] NET: Accurate packet scheduling for ATM/ADSL

2006-06-14 Thread Jesper Dangaard Brouer
The Linux traffic's control engine inaccurately calculates transmission times for packets sent over ADSL links. For some packet sizes the error rises to over 50%. This occurs because ADSL uses ATM as its link layer transport, and ATM transmits packets in fixed sized 53 byte cells. The

Re: [PATCH 0/2] NET: Accurate packet scheduling for ATM/ADSL

2006-06-14 Thread Jesper Dangaard Brouer
layer. On Wed, 2006-14-06 at 11:40 +0200, Jesper Dangaard Brouer wrote: The Linux traffic's control engine inaccurately calculates transmission times for packets sent over ADSL links. For some packet sizes the error rises to over 50%. This occurs because ADSL uses ATM as its link

Re: [PATCH 2/2] NET: Accurate packet scheduling for ATM/ADSL (userspace)

2006-06-14 Thread Jesper Dangaard Brouer
On Wed, 2006-06-14 at 11:57 +0100, Alan Cox wrote: Ar Mer, 2006-06-14 am 11:40 +0200, ysgrifennodd Jesper Dangaard Brouer: option to calculate traffic transmission times (rate table) over all ATM links, including ADSL, with perfect accuracy. The other problem I see with this code

Re: [PATCH 0/2] NET: Accurate packet scheduling for ATM/ADSL

2006-06-14 Thread Jesper Dangaard Brouer
On Wed, 2006-06-14 at 10:27 -0400, Phillip Susi wrote: Jesper Dangaard Brouer wrote: The Linux traffic's control engine inaccurately calculates transmission times for packets sent over ADSL links. For some packet sizes the error rises to over 50%. This occurs because ADSL uses ATM

Re: Network performance degradation from 2.6.11.12 to 2.6.16.20

2006-06-19 Thread Jesper Dangaard Brouer
Harry Edmon [EMAIL PROTECTED] wrote: I have a system with a strange network performance degradation from 2.6.11.12 to most recent kernels including 2.6.16.20 and 2.6.17-rc6. The system is has Dual single core Xeons with hyperthreading on. cut Hi Harry Can you check which high-res

Re: Network performance degradation from 2.6.11.12 to 2.6.16.20

2006-06-19 Thread Jesper Dangaard Brouer
On Mon, 19 Jun 2006, Andi Kleen wrote: If you use pmtmr try to reboot with kernel option clock=tsc. That's dangerous advice - when the system choses not to use TSC it often has a reason. Sorry, it was not a general advice, just something to try out. It really solved my network

Re: [PATCH 2/2] NET: Accurate packet scheduling for ATM/ADSL (userspace)

2006-06-19 Thread Jesper Dangaard Brouer
On Thu, 15 Jun 2006, jamal wrote: On Thu, 2006-15-06 at 10:47 +1000, Russell Stuart wrote: On Wed, 2006-06-14 at 11:57 +0100, Alan Cox wrote: The other problem I see with this code is it is very tightly tied to ATM cell sizes, not to solving the generic question of packetisation. Others

Re: [PATCH 0/2] NET: Accurate packet scheduling for ATM/ADSL

2006-06-20 Thread Jesper Dangaard Brouer
On Mon, 2006-06-19 at 22:35 -0700, Chris Wedgwood wrote: On Wed, Jun 14, 2006 at 11:40:04AM +0200, Jesper Dangaard Brouer wrote: The Linux traffic's control engine inaccurately calculates transmission times for packets sent over ADSL links. For some packet sizes the error rises to over

Re: [PATCH REPOST 1/2] NET: Accurate packet scheduling for ATM/ADSL (kernel)

2006-10-24 Thread Jesper Dangaard Brouer
On Tue, 24 Oct 2006, Patrick McHardy wrote: Russell Stuart wrote: On Mon, 2006-10-23 at 14:39 +0200, Patrick McHardy wrote: The implementation may be different, but the intention and the result is the same. I probably would mind less if it wouldn't affect userspace compatibility, but we

Re: Network performance degradation from 2.6.11.12 to 2.6.16.20

2006-07-04 Thread Jesper Dangaard Brouer
On Mon, 26 Jun 2006, Andi Kleen wrote: I encountered the same problem on a dual core opteron equipped with a broadcom NIC (tg3) under 2.4. It could receive 1 Mpps when using TSC as the clock source, but the time jumped back and forth, so I changed it to 'notsc', then the performance dropped

Re: Network performance degradation from 2.6.11.12 to 2.6.16.20

2006-07-10 Thread Jesper Dangaard Brouer
On Tue, 4 Jul 2006, Andi Kleen wrote: On Tuesday 04 July 2006 13:41, Jesper Dangaard Brouer wrote: Actually the change happens between kernel version 2.6.15 and 2.6.16. The timestamp optimizations are older. Don't remember the exact release, but earlier 2.6. What I'm saying

Re: [PATCH 0/2] NET: Accurate packet scheduling for ATM/ADSL (RTAB BUG)

2006-07-19 Thread Jesper Dangaard Brouer
Russell Stuart wrote: - As it stands, it doesn't help the qdiscs that use RTAB. So unless he proposes to remove RTAB entirely the ATM patch as it will still have to go in. Here is a very important point here: The RTAB (rate-table) in the kernel is NOT aligned, this is the ONLY reason

[PATCH 2/2]: [NET_SCHED]: Making rate table lookups more flexible.

2007-08-31 Thread Jesper Dangaard Brouer
commit ac093f5c2f1160ece72a6fef5c779c1892fc3152 Author: Jesper Dangaard Brouer [EMAIL PROTECTED] Date: Fri Aug 31 11:53:35 2007 +0200 [NET_SCHED]: Making rate table lookups more flexible. Extend the tc_ratespec struct, with two parameters: 1) cell_align that allow adjusting

[PATCH 1/2]: [NET_SCHED]: Make all rate based scheduler work with TSO.

2007-08-31 Thread Jesper Dangaard Brouer
commit 6fdc0f061be94f5e297650961360fb7a9d1cc85d Author: Jesper Dangaard Brouer [EMAIL PROTECTED] Date: Thu Aug 30 17:53:42 2007 +0200 [NET_SCHED]: Make all rate based scheduler work with TSO. Change L2T (length to time) macros, in all rate based schedulers, to call a common

Re: [PATCH 2/2]: [NET_SCHED]: Making rate table lookups more flexible.

2007-09-01 Thread Jesper Dangaard Brouer
On Sat, 1 Sep 2007, Patrick McHardy wrote: Jesper Dangaard Brouer wrote: [NET_SCHED]: Making rate table lookups more flexible. Extend the tc_ratespec struct, with two parameters: 1) cell_align that allow adjusting the alignment of the rate table. 2) overhead that allow adding a packet

Re: [PATCH 1/2]: [NET_SCHED]: Make all rate based scheduler work with TSO.

2007-09-01 Thread Jesper Dangaard Brouer
On Sat, 1 Sep 2007, Patrick McHardy wrote: Jesper Dangaard Brouer wrote: commit 6fdc0f061be94f5e297650961360fb7a9d1cc85d Author: Jesper Dangaard Brouer [EMAIL PROTECTED] Date: Thu Aug 30 17:53:42 2007 +0200 [NET_SCHED]: Make all rate based scheduler work with TSO. Change L2T (length

Re: [PATCH 2/2]: [NET_SCHED]: Making rate table lookups more flexible.

2007-09-02 Thread Jesper Dangaard Brouer
On Sun, 2 Sep 2007, Patrick McHardy wrote: Jesper Dangaard Brouer wrote: On Sat, 1 Sep 2007, Patrick McHardy wrote: This is not a ATM/ADSL only patch. This patch simply adds more flexibility to the rate tables. Afterwards we can start the discussion about how to use this new flexibility

Re: [PATCH 2/2]: [NET_SCHED]: Making rate table lookups more flexible.

2007-09-03 Thread Jesper Dangaard Brouer
On Sun, 2007-09-02 at 23:16 +0200, Patrick McHardy wrote: Jesper Dangaard Brouer wrote: On Sun, 2 Sep 2007, Patrick McHardy wrote: Lets focus on the general case, where the functionality actually is needed right away. In the general case: - The rate table needs to be aligned

Re: [PATCH 1/2]: [NET_SCHED]: Make all rate based scheduler work with TSO.

2007-09-05 Thread Jesper Dangaard Brouer
On Tue, 2007-09-04 at 13:40 -0400, Bill Fink wrote: On Tue, 04 Sep 2007, Patrick McHardy wrote: Bill Fink wrote: On Sat, 1 Sep 2007, Jesper Dangaard Brouer wrote: Yes, you need to specify the MTU on the command line for jumbo frames. Thanks! Works much better now, although

Re: [PATCH 2/2]: [NET_SCHED]: Making rate table lookups more flexible.

2007-09-05 Thread Jesper Dangaard Brouer
On Tue, 2007-09-04 at 18:25 +0200, Patrick McHardy wrote: Jesper Dangaard Brouer wrote: On Sun, 2007-09-02 at 23:16 +0200, Patrick McHardy wrote: Jesper Dangaard Brouer wrote: On Sun, 2 Sep 2007, Patrick McHardy wrote: Lets focus on the general case, where the functionality actually

[PATCH 0/6] NET_SCHED: Rate table fixes

2007-09-12 Thread Jesper Dangaard Brouer
This set of patches, aim at fixing an issue with the rate table used by the rate based schedulers. Currently we use the lower-boundry value, which result in under-estimating the actual bandwidth usage. The patches will change this to use the upper-boundry L2T (length to time) value. The

[PATCH 1/6] [NET_SCHED]: Cleanup L2T macros and handle oversized packets

2007-09-12 Thread Jesper Dangaard Brouer
commit a28343c933f6cfc3df1be86e0ebe8d99fa8d5f77 Author: Jesper Dangaard Brouer [EMAIL PROTECTED] Date: Wed Sep 12 10:01:00 2007 +0200 [NET_SCHED]: Cleanup L2T macros and handle oversized packets Change L2T (length to time) macros, in all rate based schedulers, to call a common

[PATCH 3/6] [IPROUTE2]: Update pkt_sched.h (to resemble the kernel one)

2007-09-12 Thread Jesper Dangaard Brouer
commit ef065a43b8900fbc0763eac0fa0a9a8a00c8aaa2 Author: Jesper Dangaard Brouer [EMAIL PROTECTED] Date: Tue Sep 11 16:17:46 2007 +0200 [IPROUTE2]: Update pkt_sched.h (to resemble the kernel one) Extend the tc_ratespec struct, with two parameters: 1) cell_align that allow

[PATCH 4/6] [IPROUTE2]: Overhead calculation is now done in the kernel

2007-09-12 Thread Jesper Dangaard Brouer
commit 07a74a2613440fc1a68d0faa7235ed7027532d78 Author: Jesper Dangaard Brouer [EMAIL PROTECTED] Date: Tue Sep 11 16:59:58 2007 +0200 [IPROUTE2]: Overhead calculation is now done in the kernel. The only current user is HTB. HTB overhead argument is now passed on to the kernel

[PATCH 5/6] [IPROUTE2]: Cleanup: tc_calc_rtable()

2007-09-12 Thread Jesper Dangaard Brouer
commit e3bad6e344303fec9916d1420aade98a2e6c79cc Author: Jesper Dangaard Brouer [EMAIL PROTECTED] Date: Wed Sep 5 10:47:47 2007 +0200 [IPROUTE2]: Cleanup: tc_calc_rtable(). Change tc_calc_rtable() to take a tc_ratespec struct as an argument. (cell_log still needs to be passed

[PATCH 6/6] [IPROUTE2]: Change the rate table calc of transmit cost to use upper bound value

2007-09-12 Thread Jesper Dangaard Brouer
commit 2e3edbef7913ac43899c8258ee59d9032778cee1 Author: Jesper Dangaard Brouer [EMAIL PROTECTED] Date: Wed Sep 5 15:24:51 2007 +0200 [IPROUTE2]: Change the rate table calc of transmit cost to use upper bound value. Patrick McHardy, Cite: 'its better to overestimate than

Re: [PATCH 4/6] [IPROUTE2]: Overhead calculation is now done in the kernel

2007-09-12 Thread Jesper Dangaard Brouer
On Wed, 2007-09-12 at 13:05 +0200, Stephen Hemminger wrote: How is this binary compatable with older kernels? It will be binary compatable, as I use/rename some unused variables in struct tc_ratespec. -- Med venlig hilsen / Best regards Jesper Brouer ComX Networks A/S Linux Network

Protocol handler for Marvell DSA EtherType packets

2008-02-01 Thread Jesper Dangaard Brouer
Hi Netdev I writing a new protocol handler using dev_add_pack(). (For a Marvell switch chip handling DSA (Distributed Switch Architecture) Ethertype packets). My protocol handler works and I get the skb. But I want to remove the DSA Headers and send the packet back for normal processing on a

Re: Protocol handler for Marvell DSA EtherType packets

2008-02-11 Thread Jesper Dangaard Brouer
://adsl-optimizer.dk LinkedIn: http://www.linkedin.com/in/brouer On Fri, 2008-02-01 at 14:28 +0100, Jesper Dangaard Brouer wrote: Hi Netdev I writing a new protocol handler using dev_add_pack(). (For a Marvell switch chip handling DSA (Distributed Switch Architecture) Ethertype packets). My

Re: [PATCH v2 net-next] net: sched: run ingress qdisc without locks

2015-05-04 Thread Jesper Dangaard Brouer
On Sun, 03 May 2015 22:12:43 -0700 Alexei Starovoitov a...@plumgrid.com wrote: On 5/3/15 8:42 AM, Jesper Dangaard Brouer wrote: I was actually expecting to see a higher performance boost. improvement diff = -2.85 ns ... The patch is removing two atomic operations, spin_{un,}lock

Re: [PATCH 1/7] net: refactor __netif_receive_skb_core

2015-04-15 Thread Jesper Dangaard Brouer
, as the result depend on the compiler version... but these kind of change do affect performance, and should be tested/measured. - -- Best regards, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http://www.linkedin.com/in/brouer

Re: Packet capturing performance

2015-05-21 Thread Jesper Dangaard Brouer
regards, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http://www.linkedin.com/in/brouer -- To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to majord...@vger.kernel.org More

Re: [net-next PATCH 06/10] pktgen: new pktgen helper functions for samples scripts

2015-05-21 Thread Jesper Dangaard Brouer
On Wed, 20 May 2015 14:23:17 -0700 Cong Wang cw...@twopensource.com wrote: On Tue, May 19, 2015 at 2:36 PM, Jesper Dangaard Brouer bro...@redhat.com wrote: + +# More generic replacement for pgset(), that does not depend on global +# variable for proc file. +function proc_cmd

Re: [RFC PATCH] slub: RFC: Improving SLUB performance with 38% on NO-PREEMPT

2015-06-08 Thread Jesper Dangaard Brouer
On Mon, 8 Jun 2015 04:39:38 -0500 (CDT) Christoph Lameter c...@linux.com wrote: On Mon, 8 Jun 2015, Jesper Dangaard Brouer wrote: My real question is if disabling local interrupts is enough to avoid this? Yes the initial release of slub used interrupt disable in the fast paths. Thanks

Re: [RFC PATCH] slub: RFC: Improving SLUB performance with 38% on NO-PREEMPT

2015-06-08 Thread Jesper Dangaard Brouer
On Thu, 04 Jun 2015 19:37:57 -0700 Eric Dumazet eric.duma...@gmail.com wrote: On Thu, 2015-06-04 at 12:31 +0200, Jesper Dangaard Brouer wrote: This patch improves performance of SLUB allocator fastpath with 38% by avoiding the call to this_cpu_cmpxchg_double() for NO-PREEMPT kernels

[RFC PATCH] slub: RFC: Improving SLUB performance with 38% on NO-PREEMPT

2015-06-04 Thread Jesper Dangaard Brouer
This patch improves performance of SLUB allocator fastpath with 38% by avoiding the call to this_cpu_cmpxchg_double() for NO-PREEMPT kernels. Reviewers please point out why this change is wrong, as such a large improvement should not be possible ;-) My primarily motivation for this patch is to

Re: [PATCH v2 00/22] FUJITSU Extended Socket network device driver

2015-06-24 Thread Jesper Dangaard Brouer
100644 drivers/net/fjes/fjes_hw.h create mode 100644 drivers/net/fjes/fjes_main.c create mode 100644 drivers/net/fjes/fjes_regs.h -- Best regards, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http

Re: [PATCH 7/7] slub: initial bulk free implementation

2015-06-16 Thread Jesper Dangaard Brouer
this_cpu_cmpxchg_double() costs 9 cycles. -- Best regards, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http://www.linkedin.com/in/brouer p.s. for comparison[1] a function call cost is 5-6 cycles, and a function pointer

[PATCH 7/7] slub: initial bulk free implementation

2015-06-15 Thread Jesper Dangaard Brouer
This implements SLUB specific kmem_cache_free_bulk(). SLUB allocator now both have bulk alloc and free implemented. Play nice and reenable local IRQs while calling slowpath. Signed-off-by: Jesper Dangaard Brouer bro...@redhat.com --- mm/slub.c | 32 +++- 1 file

[PATCH 0/7] slub: bulk alloc and free for slub allocator

2015-06-15 Thread Jesper Dangaard Brouer
://github.com/netoptimizer/prototype-kernel/blob/master/kernel/mm/slab_bulk_test01.c [2] http://workshop.netfilter.org/2015/ --- Christoph Lameter (2): slab: infrastructure for bulk object allocation and freeing slub bulk alloc: extract objects from the per cpu slab Jesper Dangaard Brouer

[PATCH 2/7] slub bulk alloc: extract objects from the per cpu slab

2015-06-15 Thread Jesper Dangaard Brouer
objects as possible in the fast way and then fall back to the generic implementation for the rest of the objects. Signed-off-by: Christoph Lameter c...@linux.com Cc: Jesper Dangaard Brouer bro...@redhat.com Cc: Pekka Enberg penb...@kernel.org Cc: David Rientjes rient...@google.com Cc: Joonsoo Kim

[PATCH 1/7] slab: infrastructure for bulk object allocation and freeing

2015-06-15 Thread Jesper Dangaard Brouer
...@linux.com Cc: Jesper Dangaard Brouer bro...@redhat.com Cc: Pekka Enberg penb...@kernel.org Cc: David Rientjes rient...@google.com Cc: Joonsoo Kim iamjoonsoo@lge.com Signed-off-by: Andrew Morton a...@linux-foundation.org --- include/linux/slab.h | 10 ++ mm/slab.c| 13

[PATCH 4/7] slub: fix error path bug in kmem_cache_alloc_bulk

2015-06-15 Thread Jesper Dangaard Brouer
kmem_cache_alloc_bulk() function. Signed-off-by: Jesper Dangaard Brouer bro...@redhat.com --- mm/slub.c | 21 - 1 file changed, 16 insertions(+), 5 deletions(-) diff --git a/mm/slub.c b/mm/slub.c index 753f88bd8b40..d10de5a33c03 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -2760,24

[PATCH 3/7] slub: reduce indention level in kmem_cache_alloc_bulk()

2015-06-15 Thread Jesper Dangaard Brouer
Use kernel early return style to reduce indention level, by testing for kmem_cache_debug() and fallback to none-optimized bulking via __kmem_cache_alloc_bulk(). This also make it easier to fix a bug in the current implementation, in the next patch. Signed-off-by: Jesper Dangaard Brouer bro

[PATCH 6/7] slub: improve bulk alloc strategy

2015-06-15 Thread Jesper Dangaard Brouer
and expensive (37 cycles) local_irq_{save,restore}. For now, be happy calling __slab_alloc() this lower icache impact of this func and I don't have to worry about correctness. Signed-off-by: Jesper Dangaard Brouer bro...@redhat.com --- mm/slub.c | 27 +-- 1 file changed, 17

[PATCH 5/7] slub: kmem_cache_alloc_bulk() move clearing outside IRQ disabled section

2015-06-15 Thread Jesper Dangaard Brouer
Move clearing of objects outside IRQ disabled section, to minimize time spend with local IRQs off. Signed-off-by: Jesper Dangaard Brouer bro...@redhat.com --- mm/slub.c | 11 --- 1 file changed, 8 insertions(+), 3 deletions(-) diff --git a/mm/slub.c b/mm/slub.c index d10de5a33c03

Re: [PATCH 6/7] slub: improve bulk alloc strategy

2015-06-17 Thread Jesper Dangaard Brouer
On Tue, 16 Jun 2015 14:53:36 -0700 Andrew Morton a...@linux-foundation.org wrote: On Mon, 15 Jun 2015 17:52:46 +0200 Jesper Dangaard Brouer bro...@redhat.com wrote: [...] + /* Invoke slow path one time, then retry fastpath +* as side-effect have

Re: [PATCH 2/7] slub bulk alloc: extract objects from the per cpu slab

2015-06-17 Thread Jesper Dangaard Brouer
On Tue, 16 Jun 2015 14:48:40 -0700 Andrew Morton a...@linux-foundation.org wrote: On Mon, 15 Jun 2015 17:52:07 +0200 Jesper Dangaard Brouer bro...@redhat.com wrote: From: Christoph Lameter c...@linux.com [NOTICE: Already in AKPM's quilt-queue] First piece: acceleration

Re: [PATCH 4/7] slub: fix error path bug in kmem_cache_alloc_bulk

2015-06-17 Thread Jesper Dangaard Brouer
On Tue, 16 Jun 2015 14:51:09 -0700 Andrew Morton a...@linux-foundation.org wrote: On Mon, 15 Jun 2015 17:52:26 +0200 Jesper Dangaard Brouer bro...@redhat.com wrote: The current kmem_cache/SLAB bulking API need to release all objects in case the layer cannot satisfy the full request

Re: [PATCH 7/7] slub: initial bulk free implementation

2015-06-16 Thread Jesper Dangaard Brouer
On Tue, 16 Jun 2015 16:23:28 +0900 Joonsoo Kim iamjoonsoo@lge.com wrote: On Mon, Jun 15, 2015 at 05:52:56PM +0200, Jesper Dangaard Brouer wrote: This implements SLUB specific kmem_cache_free_bulk(). SLUB allocator now both have bulk alloc and free implemented. Play nice

Re: [PATCH 7/7] slub: initial bulk free implementation

2015-06-16 Thread Jesper Dangaard Brouer
the results. -- Best regards, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http://www.linkedin.com/in/brouer -- To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to majord

Re: [PATCH 7/7] slub: initial bulk free implementation

2015-06-16 Thread Jesper Dangaard Brouer
On Tue, 16 Jun 2015 10:21:10 +0200 Jesper Dangaard Brouer bro...@redhat.com wrote: On Tue, 16 Jun 2015 16:28:06 +0900 Joonsoo Kim iamjoonsoo@lge.com wrote: Is this really better than just calling __kmem_cache_free_bulk()? Yes, as can be seen by cover-letter, but my cover-letter does

Re: [PATCH 7/7] slub: initial bulk free implementation

2015-06-16 Thread Jesper Dangaard Brouer
On Mon, 15 Jun 2015 11:34:44 -0500 (CDT) Christoph Lameter c...@linux.com wrote: On Mon, 15 Jun 2015, Jesper Dangaard Brouer wrote: + for (i = 0; i size; i++) { + void *object = p[i]; + + if (unlikely(!object)) + continue; // HOW ABOUT BUG_ON

Re: [PATCHv3] pktgen: Convert return type of process_ipsec to bool

2015-05-28 Thread Jesper Dangaard Brouer
the main Linux Kernel Mailing List (linux-ker...@vger.kernel.org) with a trivial patch like this. Sending it to the network developers should be enough (netdev@vger.kernel.org). -- Best regards, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv

Re: [net-next PATCH 09/10] pktgen: add sample script pktgen_sample03_burst_single_flow.sh

2015-05-21 Thread Jesper Dangaard Brouer
On Wed, 20 May 2015 14:33:42 -0700 Cong Wang cw...@twopensource.com wrote: On Tue, May 19, 2015 at 2:36 PM, Jesper Dangaard Brouer bro...@redhat.com wrote: +# Threads are specified with parameter -t value in $THREADS +for ((thread = 0; thread $THREADS; thread++)); do +dev=${DEV

[net-next PATCH v2 00/10] pktgen: cleanups and introducing new samples/pktgen scripts

2015-05-21 Thread Jesper Dangaard Brouer
/netoptimizer/network-testing/tree/master/pktgen --- Jesper Dangaard Brouer (10): pktgen: add benchmark script pktgen_bench_xmit_mode_netif_receive.sh pktgen: add sample script pktgen_sample03_burst_single_flow.sh pktgen: add sample script pktgen_sample02_multiqueue.sh pktgen

[net-next PATCH v3 05/10] pktgen: make /proc/net/pktgen/pgctrl report fail on invalid input

2015-05-21 Thread Jesper Dangaard Brouer
scripting a way of detecting this. Also bump version tag to 2.75, because (1) reading /proc/net/pktgen/pgctrl output this version number which would allow to detect this small semantic change, and (2) because the pktgen version tag have not been updated since 2010. Signed-off-by: Jesper Dangaard Brouer

[net-next PATCH v3 03/10] pktgen: doc were missing several config options

2015-05-21 Thread Jesper Dangaard Brouer
vlan_p svlan_cfi svlan_id svlan_p Signed-off-by: Jesper Dangaard Brouer bro...@redhat.com --- Documentation/networking/pktgen.txt | 25 - 1 files changed, 24 insertions(+), 1 deletions(-) diff --git a/Documentation/networking/pktgen.txt b/Documentation/networking

[net-next PATCH v3 04/10] pktgen: document ability to add same device to several threads

2015-05-21 Thread Jesper Dangaard Brouer
: e6fce5b916cd7 (pktgen: multiqueue etc.) Signed-off-by: Jesper Dangaard Brouer bro...@redhat.com Acked-by: Alexei Starovoitov a...@plumgrid.com --- Documentation/networking/pktgen.txt | 77 +++ 1 files changed, 50 insertions(+), 27 deletions(-) diff --git

[net-next PATCH v2 01/10] pktgen: remove obsolete max_before_softirq from pktgen doc

2015-05-21 Thread Jesper Dangaard Brouer
And cleanup some whitespaces in pktgen.txt. Signed-off-by: Jesper Dangaard Brouer bro...@redhat.com --- 0 files changed, 0 insertions(+), 0 deletions(-) diff --git a/Documentation/networking/pktgen.txt b/Documentation/networking/pktgen.txt index 747facc..62f5ebb 100644 --- a/Documentation

[net-next PATCH v3 06/10] pktgen: new pktgen helper functions for samples scripts

2015-05-21 Thread Jesper Dangaard Brouer
/netoptimizer/network-testing/tree/master/pktgen Signed-off-by: Jesper Dangaard Brouer bro...@redhat.com --- v2: - remove useless cat's, but keep them in old pgset() - add comment due to pgctrl, cannot use exit code $? from grep samples/pktgen/README.rst| 43 +++ samples

[net-next PATCH v3 10/10] pktgen: add benchmark script pktgen_bench_xmit_mode_netif_receive.sh

2015-05-21 Thread Jesper Dangaard Brouer
by Alexei Starovoitov. This script don't really need any hardware. This is achieved via the recently introduced stack inject feature xmit_mode netif_receive. See commit 62f64aed622b6 (pktgen: introduce xmit_mode 'start_xmit|netif_receive'). Signed-off-by: Jesper Dangaard Brouer bro...@redhat.com

[net-next PATCH v2 02/10] pktgen: adjust spacing in proc file interface output

2015-05-21 Thread Jesper Dangaard Brouer
Too many spaces were introduced in commit 63adc6fb8ac0 (pktgen: cleanup checkpatch warnings), thus misaligning src_min: to other columns. Fixes: 63adc6fb8ac0 (pktgen: cleanup checkpatch warnings) Signed-off-by: Jesper Dangaard Brouer bro...@redhat.com --- 0 files changed, 0 insertions(+), 0

Re: [net-next PATCH v2 01/10] pktgen: remove obsolete max_before_softirq from pktgen doc

2015-05-21 Thread Jesper Dangaard Brouer
On Thu, 21 May 2015 12:11:42 +0200 Jesper Dangaard Brouer bro...@redhat.com wrote: --- 0 files changed, 0 insertions(+), 0 deletions(-) Aborting V2 stg-mail sending, because it's not generating the correct diff stat. This is a bug in stg if not in the root of the git dir. -- Best

[net-next PATCH v3 00/10] pktgen: cleanups and introducing new samples/pktgen scripts

2015-05-21 Thread Jesper Dangaard Brouer
bursting of SKBs -v : ($VERBOSE) verbose -x : ($DEBUG) debug These scripts are borrowed from: https://github.com/netoptimizer/network-testing/tree/master/pktgen --- Jesper Dangaard Brouer (10): pktgen: add benchmark script pktgen_bench_xmit_mode_netif_receive.sh pktgen: add

[net-next PATCH v3 02/10] pktgen: adjust spacing in proc file interface output

2015-05-21 Thread Jesper Dangaard Brouer
Too many spaces were introduced in commit 63adc6fb8ac0 (pktgen: cleanup checkpatch warnings), thus misaligning src_min: to other columns. Fixes: 63adc6fb8ac0 (pktgen: cleanup checkpatch warnings) Signed-off-by: Jesper Dangaard Brouer bro...@redhat.com --- net/core/pktgen.c |2 +- 1 files

[net-next PATCH v3 01/10] pktgen: remove obsolete max_before_softirq from pktgen doc

2015-05-21 Thread Jesper Dangaard Brouer
And cleanup some whitespaces in pktgen.txt. Signed-off-by: Jesper Dangaard Brouer bro...@redhat.com --- Documentation/networking/pktgen.txt | 18 -- 1 files changed, 8 insertions(+), 10 deletions(-) diff --git a/Documentation/networking/pktgen.txt b/Documentation/networking

[net-next PATCH v3 07/10] pktgen: add sample script pktgen_sample01_simple.sh

2015-05-21 Thread Jesper Dangaard Brouer
: ($DEV) output interface/device (required) -s : ($PKT_SIZE) packet size -d : ($DEST_IP) destination IP -m : ($DST_MAC) destination MAC-addr -c : ($SKB_CLONE) SKB clones send before alloc new SKB -v : ($VERBOSE) verbose -x : ($DEBUG) debug Signed-off-by: Jesper Dangaard

[net-next PATCH v3 09/10] pktgen: add sample script pktgen_sample03_burst_single_flow.sh

2015-05-21 Thread Jesper Dangaard Brouer
38b2cf2982dc73 (net: pktgen: packet bursting via skb-xmit_more). [1] http://netoptimizer.blogspot.dk/2014/06/pktgen-for-network-overload-testing.html [2] http://netoptimizer.blogspot.dk/2014/10/unlocked-10gbps-tx-wirespeed-smallest.html Signed-off-by: Jesper Dangaard Brouer bro...@redhat.com --- v2

[net-next PATCH v3 08/10] pktgen: add sample script pktgen_sample02_multiqueue.sh

2015-05-21 Thread Jesper Dangaard Brouer
: ($VERBOSE) verbose -x : ($DEBUG) debug Removing pktgen.conf-2-1 and pktgen.conf-2-2 as these examples should be covered now. Signed-off-by: Jesper Dangaard Brouer bro...@redhat.com --- Documentation/networking/pktgen.txt |2 - samples/pktgen/pktgen.conf-2-1

Re: [net-next PATCH 1/3] net: make default tx_queue_len configurable

2015-08-13 Thread Jesper Dangaard Brouer
, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http://www.linkedin.com/in/brouer -- To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to majord...@vger.kernel.org More majordomo

Re: [PATCH 0/2] net: introduce IFF_NO_QUEUE as successor of zero tx_queue_len

2015-08-17 Thread Jesper Dangaard Brouer
On Mon, 17 Aug 2015 06:44:51 -0700 Eric Dumazet eric.duma...@gmail.com wrote: On Mon, 2015-08-17 at 08:51 +0200, Jesper Dangaard Brouer wrote: On Fri, 14 Aug 2015 10:41:53 +0200 Phil Sutter p...@nwl.cc wrote: On Thu, Aug 13, 2015 at 12:11:57PM -0700, Stephen Hemminger wrote

Re: [net-next PATCH v2 3/3] net: sched: fall back to noqueue when removing root qdisc

2015-08-23 Thread Jesper Dangaard Brouer
to assign noqueue (to any device). This makes the action explicit for the user, instead of being a side-effect of removing a qdisc. (and does not break backward compat) -- Best regards, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv-analyzer.org

Re: [net-next PATCH v2 3/3] net: sched: fall back to noqueue when removing root qdisc

2015-08-23 Thread Jesper Dangaard Brouer
On Sun, 23 Aug 2015 20:44:42 +0200 Jesper Dangaard Brouer bro...@redhat.com wrote: On Sat, 22 Aug 2015 02:20:56 +0200 Phil Sutter p...@nwl.cc wrote: When removing the root qdisc, the interface should fall back to noqueue as the 'real' minimal qdisc instead of the default one. I worry

Re: [PATCH 2/2] net: sch_generic: react upon IFF_NO_QUEUE flag

2015-08-17 Thread Jesper Dangaard Brouer
On Thu, 13 Aug 2015 19:01:07 +0200 Phil Sutter p...@nwl.cc wrote: Handle IFF_NO_QUEUE as alternative to tx_queue_len being zero. Signed-off-by: Phil Sutter p...@nwl.cc Acked-by: Jesper Dangaard Brouer bro...@redhat.com -- To unsubscribe from this list: send the line unsubscribe netdev

Re: [PATCH 0/2] net: introduce IFF_NO_QUEUE as successor of zero tx_queue_len

2015-08-17 Thread Jesper Dangaard Brouer
is the only solution *not* breaking existing scripts. So if you see the rub, I would like to know where exactly it is. I agree with Phil. AFAIC see this approach does not break existing scripts. Acked-by: Jesper Dangaard Brouer bro...@redhat.com -- Best regards, Jesper Dangaard Brouer MSc.CS, Sr

Re: [PATCH 1/2] net: declare new net_device priv_flag IFF_NO_QUEUE

2015-08-17 Thread Jesper Dangaard Brouer
Dangaard Brouer bro...@redhat.com -- Best regards, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http://www.linkedin.com/in/brouer -- To unsubscribe from this list: send the line unsubscribe netdev in the body

Re: [net-next PATCH 1/3] net: make default tx_queue_len configurable

2015-08-11 Thread Jesper Dangaard Brouer
to the default value (e.g. remove setting it to zero). -- Best regards, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http://www.linkedin.com/in/brouer -- To unsubscribe from this list: send the line unsubscribe netdev

Re: [PATCH 0/2] net: introduce IFF_NO_QUEUE as successor of zero tx_queue_len

2015-08-13 Thread Jesper Dangaard Brouer
it. -- Best regards, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http://www.linkedin.com/in/brouer -- To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to majord...@vger.kernel.org More

[PATCH 2/4] net: bulk free SKBs that were delay free'ed due to IRQ context

2015-10-23 Thread Jesper Dangaard Brouer
needed. This due to netpoll can call from IRQ context. Signed-off-by: Alexander Duyck <alexander.h.du...@redhat.com> Signed-off-by: Jesper Dangaard Brouer <bro...@redhat.com> --- include/linux/skbuff.h |1 + net/core/dev.c |8 +++- net/core/skbuff.c |8 +

[PATCH 4/4] net: bulk alloc and reuse of SKBs in NAPI context

2015-10-23 Thread Jesper Dangaard Brouer
t (normal hidden by prefetch) * In case RX queue is not full, alloc and free more SKBs than needed More testing is needed with more real life benchmarks. Joint work with Alexander Duyck. Signed-off-by: Jesper Dangaard Brouer <bro...@redhat.com> Signed-off-by: Alexander Duyck <alexander.

[PATCH 0/4] net: mitigating kmem_cache slowpath for network stack in NAPI context

2015-10-23 Thread Jesper Dangaard Brouer
ey change virt_to_head_page(), had to apply patches manually from http://ozlabs.org/~akpm/mmotm/ (stamp-2015-10-20-16-33) as AKPM made several small fixes. --- Jesper Dangaard Brouer (4): net: bulk free infrastructure for NAPI context, use napi_consume_skb net: bulk free SKBs that were

[PATCH 3/4] ixgbe: bulk free SKBs during TX completion cleanup cycle

2015-10-23 Thread Jesper Dangaard Brouer
1.1-4)) Joint work with Alexander Duyck. Signed-off-by: Alexander Duyck <alexander.h.du...@redhat.com> Signed-off-by: Jesper Dangaard Brouer <bro...@redhat.com> --- drivers/net/ethernet/intel/ixgbe/ixgbe_main.c |6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --gi

[PATCH 1/4] net: bulk free infrastructure for NAPI context, use napi_consume_skb

2015-10-23 Thread Jesper Dangaard Brouer
is to see if budget is 0. In that case, we need to invoke dev_consume_skb_irq(). Joint work with Alexander Duyck. Signed-off-by: Jesper Dangaard Brouer <bro...@redhat.com> Signed-off-by: Alexander Duyck <alexander.h.du...@redhat.com> --- include/linux/skbuff.h |3 ++ ne

Re: [PATCH net-next] sock: don't enable netstamp for af_unix sockets

2015-10-27 Thread Jesper Dangaard Brouer
[kernel.vmlinux] [k] ktime_get_with_offset (1/1989575*10^9)*((1.54+1.07)/100) = 13.12 nanosec On some of my slower systems, I've seen cost of just reading TSC be around 32 ns. > Also counter question: why is the netstamp code protected by a > static_key otherwise if not for trying to suppress

Re: [MM PATCH V4 6/6] slub: optimize bulk slowpath free by detached freelist

2015-10-21 Thread Jesper Dangaard Brouer
On Wed, 14 Oct 2015 14:15:25 +0900 Joonsoo Kim <iamjoonsoo@lge.com> wrote: > On Tue, Sep 29, 2015 at 05:48:26PM +0200, Jesper Dangaard Brouer wrote: > > This change focus on improving the speed of object freeing in the > > "slowpath" of kmem_cache_free_bul

Re: [MM PATCH V4.1 5/6] slub: support for bulk free with SLUB freelists

2015-10-07 Thread Jesper Dangaard Brouer
On Mon, 5 Oct 2015 21:26:39 +0200 Jesper Dangaard Brouer <bro...@redhat.com> wrote: > On Fri, 2 Oct 2015 14:50:44 -0700 Andrew Morton <a...@linux-foundation.org> > wrote: > [...] > > > Deleting the BUG altogether sounds the best solution. As long as the >

Re: [MM PATCH V4.1 5/6] slub: support for bulk free with SLUB freelists

2015-10-07 Thread Jesper Dangaard Brouer
On Tue, 6 Oct 2015 01:07:03 +0200 Jesper Dangaard Brouer <bro...@redhat.com> wrote: > (trimmed Cc list a little) > > On Mon, 5 Oct 2015 14:20:45 -0700 Andi Kleen <a...@linux.intel.com> wrote: > > > > My only problem left, is I want a perf measurement that p

Re: [MM PATCH V4.1 5/6] slub: support for bulk free with SLUB freelists

2015-10-05 Thread Jesper Dangaard Brouer
On Fri, 2 Oct 2015 14:50:44 -0700 Andrew Morton <a...@linux-foundation.org> wrote: > On Fri, 2 Oct 2015 15:40:39 +0200 Jesper Dangaard Brouer <bro...@redhat.com> > wrote: > > > > Thus, I need introducing new code like this patch and at the same time &

[RFC PATCH 2/3] net: NIC helper API for building array of skbs to free

2015-09-04 Thread Jesper Dangaard Brouer
t of these functions. Signed-off-by: Jesper Dangaard Brouer <bro...@redhat.com> --- include/linux/netdevice.h | 62 + 1 file changed, 62 insertions(+) diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 05b9a694e213..d0133e778314

[RFC PATCH 1/3] net: introduce kfree_skb_bulk() user of kmem_cache_free_bulk()

2015-09-04 Thread Jesper Dangaard Brouer
ion and use this in all call points. Signed-off-by: Jesper Dangaard Brouer <bro...@redhat.com> --- include/linux/skbuff.h |1 + net/core/skbuff.c | 87 +++- 2 files changed, 71 insertions(+), 17 deletions(-) diff --git a/include/linux/s

[RFC PATCH 3/3] ixgbe: bulk free SKBs during TX completion cleanup cycle

2015-09-04 Thread Jesper Dangaard Brouer
. Choosing bulk size 32 even-though budget usually is 64, due (1) to limit the stack usage and (2) as SLAB behind SKBs have 32 objects per slab. Signed-off-by: Jesper Dangaard Brouer <bro...@redhat.com> --- drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 13 +++-- 1 file chang

[RFC PATCH 0/3] Network stack, first user of SLAB/kmem_cache bulk free API.

2015-09-04 Thread Jesper Dangaard Brouer
ding: - unloaded netfilter modules - Sysctl settings: - net/ipv4/conf/default/rp_filter = 0 - net/ipv4/conf/all/rp_filter = 0 - (Forwarding performance is affected by early demux) - net/ipv4/ip_early_demux = 0 - net.ipv4.ip_forward = 1 - Disabled GRO on NICs - ethtool -K ixgbe3 gro off t

Re: [RFC PATCH 0/3] Network stack, first user of SLAB/kmem_cache bulk free API.

2015-09-05 Thread Jesper Dangaard Brouer
etached freelist, which allow me to return several objects with a single locked double_cmpxchg. -- Best regards, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http://www.linkedin.com/in/brouer -- To unsubscribe fro

Re: [RFC PATCH 1/3] net: introduce kfree_skb_bulk() user of kmem_cache_free_bulk()

2015-09-07 Thread Jesper Dangaard Brouer
On Fri, 4 Sep 2015 11:47:17 -0700 Tom Herbert <t...@herbertland.com> wrote: > On Fri, Sep 4, 2015 at 10:00 AM, Jesper Dangaard Brouer <bro...@redhat.com> > wrote: > > Introduce the first user of SLAB bulk free API kmem_cache_free_bulk(), > > in the net

Re: [RFC PATCH 0/3] Network stack, first user of SLAB/kmem_cache bulk free API.

2015-09-07 Thread Jesper Dangaard Brouer
ng to the next page. Thus, slub bulk alloc will give many objects belonging to the page. I'm trying to keep these objects grouped together until they can be free'ed in a bulk. -- Best regards, Jesper Dangaard Brouer MSc.CS, Sr. Network Kernel Developer at Red Hat Author of http://www.iptv-analy

Re: slab-nomerge (was Re: [git pull] device mapper changes for 4.3)

2015-09-07 Thread Jesper Dangaard Brouer
I'm working on speeding up slub to the level of slab. And it seems like I have succeeded with half-a-nanosec 2090522 pps (+2227 pps or 0.51 ns). And with "slab_nomerge" I get even high performance: * slub: bulk-free and slab_nomerge: 2121824 pps * Diff to slub: +78249 and -18.05ns --

Re: [PATCH mm] slab: implement bulking for SLAB allocator

2015-09-08 Thread Jesper Dangaard Brouer
On Tue, 8 Sep 2015 10:22:32 -0500 (CDT) Christoph Lameter <c...@linux.com> wrote: > On Tue, 8 Sep 2015, Jesper Dangaard Brouer wrote: > > > Also notice how well bulking maintains the performance when the bulk > > size increases (which is a soar spot for the slub a

  1   2   3   4   5   6   7   8   9   10   >